EDSAC Simulator - Computerphile
Vložit
- čas přidán 20. 02. 2018
- Play around with one of the first Von Neumann machines (well, in simulation!) Professor Brailsford shows us what EDSAC was capable of...
Our special thanks go to Professor Martin Campbell-Kelly for the simulator and advice.
If you want to play with the simulator, you can find it here: bit.ly/c_EDSACSim
/ computerphile
/ computer_phile
This video was filmed and edited by Sean Riley.
Computer Science at the University of Nottingham: bit.ly/nottscomputer
Computerphile is a sister project to Brady Haran's Numberphile. More at www.bradyharan.com
"You will program in the crudest and most brain-damaging form of assembly you can imagine.". Wonderful.
I enjoyed that.
... and you're gonna like it, and crying out loud "harder daddy".
*Esoteric languages entered the chat*
Compteurs système d'exploitation
It's my brain that'll need to be (re)assembled...
I wish I was studying at Nottingham... Professor Brailsford is so gifted, not only with intelligence but also with the ability to teach without being boring. I could listen to him all day.
Me too !
Don't know if the prof reads these comments, but I wish he was my computer science teacher back in the day. The university is lucky to have you professor, and we're lucky to have you on youtube. Thanks for your time.
Thanks for this. The best comments for me -- and my reason in many ways for being on Computerphile - is when people say " I didn't relly nderstand it until you explained it "
ProfDaveB Professor, I love seeing videos with you, you're an amazing storyteller! I've learned so much about computer history. Do you ever come to visit the US?
Thanks for the kind comments - always much appreciated. It's particularly rewarding for me when people say things like: "I didn't understand this topic until you explained it".
I used to come to the US twice a year until about 2013 but circumstances have changed in recent years (since "Computerphile" started) and I don't come nearly as often now.
"Thermionic" is an excellent word that needs a new life in computer science
"...in 1949, life hadn't been invented." You heard it here first, folks! -- Out of Context Digest, Issue 82.
No extraneous characters allowed, so no comments. But you're free to write whatever comments you want on the tape with a pencil!
We need more of your videos mate. Is nothing of note happening in the AI world?
Fermat tried, not enough space.
@Bagana This comment wins the internet for February. Well done :)
No! Do not write on paper tape, especially in pencil. I used to load paper tape records of long distance phone calls back in the 80's, when I worked as a mainframe computer operator. The readers were considered ancient even then, so spare parts were rare and expensive. Paper with bends, indentations from writing, or any other changes was more likely to jam in the clunky mechanical reader. I don't know if graphite flaking off from pencil marks would damage the gears or help lubricate them, but it would be foolish to experiment. The punch card readers were a lot more forgiving, so you could write on cards, but on paper tape we only wrote on the leading section that didn't run through all the gears.
William, the clay in the pencil would be abrasive. We had a rule that, even if the paper was only dirty, you duplicated it and threw the original away, after checking the new tape. That was around 1975, at Plymouth, where students used PDP-8's for most of the time. The Poly also had an ICL (0.5 MB RAM) and an IBM (1.0 MB RAM), which was the most powerful computer in the Southwest of England, and had cost £2,000,000, and more for rental of the 180 kB per platter Winchester drive and purchase of other accessories. The two computers shared a floor of a building block, plus other space for power supplies and ventilation. Both of the mainframes used card readers or Creed 2300 teleprinters for input, and paper sheets or teleprinters for output. The cards were definitely more tolerant of damage than the paper tapes. Access to the computer rooms was restricted to computer staff only, and clean room conditions were observed. Oddly, they had no monitors on the mainframes, but a couple of orange monitors for use with the PDP-8's. We also had a small analogue computer and chart plotter which was an interesting beast, but I've no idea about who built it. It could calculate and plot ballistic trajectories, simple weather data, thermionic and semiconductor device characteristics, etc.
Prof. Brailsford is the Sir Attenborough of computing.
Wow! This brings back memories, it's like time travel back to my college years! I so enjoyed this and look forward to more conversations with Brailsford!
Real programmers use a knife to engrave the bits directly into paper strips and beatbox directly onto the mercury delay line :)
Real programmers don't use any sort of tools or computers, but just compute everything in their head, because that's much faster, just like John von Neumann did.
real programmers write the dna of programmers who write the code.
Real programmers write the laws of chemistry and physics for the programmers that write the DNA of the programmers that write the code.
Human computers were around before DNA was even thought about.
Real programmers use butterflies.
The world needs more teachers like him. It's always a pleasure to watch his videos.
I love the videos that feature professor Brailsford. He’s just fantastic.
So amazing to see how computing was done years before I was born. An absolute privilege to be able to play around with a first generation computer.
"When you needed these people, they came along."
Ah, if we may borrow a variant of "the Anthropic Principle" from the physicists, the answer to that is that, if they didn't show up, then it wouldn't have happened.
So, where it did happen, then of course you had the right people at the right time there to do it. Before if you didn't have that, then it couldn't have happened and we wouldn't be talking about it. At least until the right people at the right time did eventually show up to make it happen, and give us some history to discuss.
14:00 "life hadn't been invented" wow the computer is older than life itself ! Lol. Great to see more content from prof B on this channel. He is so good at telling the history and making it interesting.
fredhair wow, I was reading through the comments while watching the video and the exact line was on when reading this comment!
I wish I had 1/1000 of his wit and charm and brains........oh well. Such a nice informative video. Thank you so much for posting.
I wrote a simulator of EDSAC II for my dissertation about 25 years ago...running the original "ROM" code and seeing it work was hugely satisfying
Link please
Congrats on 1 Mil Subscribers!
The fact that I'm looking at a video about one of the first computers on a phone that is thousands of times more powerful than it, but takes up a tiiiiny fraction of the space and power, is absolutely fascinating. Imagine where we will go.
I really like this channel.
Professor Brailsford reminds me so much of my lovely uncle (who is 77 soon) in both knowledge, ability, personality and temperament, who also has spent his life in electronics and computers, travelling the world on behalf of his university and industry. He also did high level maths, programming and development and was involved in cutting edge technology, which often nobody knew how to use. He also gave many lectures and trained young college kids to prepare them for a life in industry. In his younger days he worked as a TV engineer for Baird's in Yorkshire, England (UK).
EgoShredder That was a really interesting story :) Your uncle has definitely seen a lot. I hope he is doing well.
He's doing reasonably ok for his age and likes to keep himself occupied with projects. To add to this, he left school in the 1950s and climbed onto roofs installing TV aerials / fixing sets. If he encountered any troublesome customers being a pain, he would send all soot down the chimney, and then the customer coming out with a black face etc! He has some hilarious stories to tell from those days. He also returned to school himself aged 38 around 1979 approx to refresh his basic qualifications from the 1950s. He sat amongst 16 year olds doing their lessons and exams, while they all laughed at him for being there etc. He had the last laugh, as he soon caught them up and overtook them. He had to do this before he could take the advanced mathematics courses and exams.
I expect Prof Brailsford is also like my uncle, in that he reads circuit diagrams and computer programming code, and then corrects any design and code errors and informs the manufacturer / creators so they can update things.
EgoShredder HA, yeah that is hillarious!
I bet the 1970's and early 80's was a really exciting time to work with computers. I was born in 1977, so I just missed out on it. I do remember using computers at school when I was a little kid in the 80's, when they were using 5.25 inch floppy disks. I remember a tiny bit of the early personal computer era. But I was too young to really understand much about it. I knew computers were around and it was exciting, but I never really got to use them that much myself. I mostly just played educational games at school. We didn't have a computer at home. (They were still VERY expensive in the 80's, at least here in the US). So even though I was around at that time, I didn't get to experience much of it. Now I've started collecting vintage computers as a hobby and learning how to program them with BASIC. I wish I could have been just a little bit older so I could have done it back then!
Born in 1971 here in England (UK) and yes the late 70s onwards was a very magical time indeed, and I do not use the word magic lightly. The vast majority of every day items were still analogue or manually operated, much like they had been for decades. A few things here and there had microprocessors / logic chips, to handle minor tasks. So full sized computers were these strange new devices, that did weird and wonderful new things that nobody understood, unless they were interested or involved in science.
"...otherwise that wretched machine is going to win." 😄
Will you talk about the very clever stuff Konrad Zuse did for memory (and his VERY clever relay based adder)
Paper tape is the most efficient storage method ever invented. The data are stored in holes, and the paper, being there only to hold the holes in place, is 100% redundant.
What do you mean 100% redundant? Can't it get torn, or only get partially punched due to worn out hardware?
It's just an old computing joke. It wasn't meant to be taken seriously, sorry! :)
I hope there will be (some day in the future) a short video about the Zuse Z-1 etc. - just to get an outside look.
Professor Brailsford is spectacular!!!!
As I understand it, mercury delay lines were used in radar signal processing to remove static objects. Two pulses would be sent out in the same direction at slightly different times, with the time difference a multiple of the radio period. The earlier pulse would be sent into the delay line, and when the second pulse was received, it would be inverted and then added to the pulse from the delay line to display on the screen. That way, any object in the same position for both pulses would be masked, and only moving objects would show up.
So it was kind of being used for memory, but at the time it wasn't being recycled, so it only held data for a fraction of a second. This would not be suitable for recording radar pulses, because the analog signal would rapidly degrade due to noise in the line. However, for a digital signal, a new, clean signal can be produced every cycle, and there is no data degradation.
I'd love Professor Brailsford to talk about the Austrian Mailüfterl computer, which was to my knowledge the first fully transistor based general computer on the European continent.
sweet so many new films on old computing, like the Apollo to the moon computer which is much newer and then where we are today... Ai, etc. 😎 thanks
Thanks for showing me this emulator ... I didn't know this existed .... !
Great stuff!
I told my GCSE students about the power usage of early computers. When you tell them "they used to use about the same power as 11 average houses - so your estate" they get a bit surprised :)
For comparison: now vs then:
What's the average power consumption of google ?
Probably hundreds of MW.
Frank Schneider 3.756 GW
But they didn’t count the “cloud” we have now
It's strange to think that even how primitive it is, it could still theoretically do everything a modern computer does, given additional memory and processing speed.
Is there going to be a follow up video on the LEO I? It's especially interesting to a lot of us who work in business programming to see the granddaddy of all business hardware/applications and as I understand it was a spin-off of the EDSAC project.
Excellent topic
Gongrats on a million subs!
The mercury "memory" reminded me of a electronic calculator that used piano wire wound around and also used acoustics to "hold" data.
Wow! It's crazy just how far we have come. It's fascinating hearing about all of these janky techniques that were used to get around the limitations of technology at the time.
Awesome video! Btw, the simulator does work on WINE for anyone interested.
I am watching these videos about first computers , and getting a feel on what it would be like in that time.
I am wandering would this machine even struck me as a computer.
So how hard is it to write a program for the Edsac? I admit I know zilch about programming, but I do like old computers, and I'd like to have a go with the emulator.
Fascinating ...
So, this primal memory system reminds me myself going to the supermarket, saying and hearing the things I need to buy.
1:13 Background top right, that looks like the magnetic part inside a floppy disc, but it's HUGE!
Is it a magnetic disc or something else?
That "display" screams for a Tetris clone.
MrTridac Would be too slow for even tetris.
I worked at a Tool and Die Machine shop when I was in college during my brakes and they had a CNC that used 5 hold paper tape to load jobs onto it. This was back around 2002 or so.
You've been living in Africa at that time ?
Frank Schneider Machine shops love obsolete tech. Not so bad if they keep maintenance up, can really be a pain when you've had a couple decades of plant managers pinching pennies and trying to pass every expense onto the next guy.
so close to 1 million!!
This video resonates a bit better than the computer nostalgia one I watched before where a guy ridicules the specs of the Pentium 4. Can you imagine quoting those specs to someone for whom this was revolutionary?
How would I go about learning to use this? Sounds really interesting. Anybody have a series of courses or videos?
Million subs!! Well deserved, too! Bravo, guys. Keep it up, you'll be at 2mil in no time :)
Could you do some videos on the RISC-V architecture?
There was a similar emulator for CSIRAC's instruction set written back in the late 1990s (Windows 9x executable), though unlike this EDSAC one, it doesn't emulate the machine's speed.
I would love to see Professor Brailsford react to the kind of hardware on LinusTechTips
How many mercury memory tubes did they have? One for operating memory and another one to feed the display? Did they have still another one for the instructions?
How do you define optimal start for noughts and crosses?
If you start in the centre your opponent has a choice of 4 of the remaining 8 squares which if chosen allow you to (force a) win.
If you start at a centre of one side there are again 4 of the reamining 8 squares that if chosen allow you to (force a) win
But if you start in a corner there are 7 of the remaining 8 squares that if chosen by your oppenent allow you to (force a) win.
In other words if your opponent was to choose their reply at random if you start in the centre or the centre of an edge you could force a win in half the games, but if you start at a corner you can force a win in 7 out of 8 games!
i got this to work on Linux Arch using the latest wine 3.2
I wonder if they could have used clay records for memory. The reading head could be much lighter than writing head so you could use the record as read/write memory. To format you just flatten the disk...
* real programmers debug with oscilloscopes!
* you had ones and zeroes! - man, all we had were zeroes!
Why isn't this man Knighted?
Does anyone know anything about the diamond v computer? Thanks
12:50 Well, that gives "dial-up networking" a whole new meaning...
I believe that particular tic-tac-toe program is what the CZcamsr Ahoy decided was, according to his personal criteria, the first video game.
(Not the first computer game, or even the first computer tic-tac-toe game--but the first to have a video display, which makes it a video game, and it predates Tennis for Two and Spacewar! There's a patent for an oscilloscope-based game that is earlier, but there's no evidence that it was ever actually built.)
When Morris Wilkes was putting this machine together, and put the electronic engineer (Bill Rennick (if I'm spelling his name right)) ..and David Wheeler, along with the others working on hardware / software.. I would have liked to have been a fly on the wall when this was being worked out. Software guys who needed to know what hardware it would run on, hardware guys who needed to know what it would be doing, and Morris Wilkes answering a thousand questions a day. LOL
I've always wondered why they used paper tape and not used up film reels which would've been available and would be stronger than paper tape as the tape ran continuously in a loop.
Liked the "Mercury-pipe', great story. Thank you Professor Brailsford.
12:06 Optimal strategy is to grab the centre...actually I'd have to disagree.
If you take centre first your opponent has a choice of 4 of the remaining 8 cells where it will not lead to a forced loss, a 50% random chance of picking a non-losing cell.
If you pick one of the centre edge cells, again there are 4 cells of the remaining 8 which do not lead to a forced loss.
If you pick a corner cell, of the remaining 8 cells only 1 leads to a non-forced-loss, or an 87% random chance that a cell picked at random will lead to a forced loss.
Unless your opponent knows that *one* cell, the optimum strategy is to pick a corner cell as you are more likely to force a win!
(In case your wondering is going on, the only cell that prevents a forced loss in reply to a corner opening is to take the centre cell; incidentally the centre cell reply is also one of the four cells which do not lead to a forced loss if the first player starts with a side centre cell. If the first player takes the centre cell, the non-loss forcing replies are the corner cells.
The optimum strategy first move actually only exists for the second player: if the centre is free take it otherwise take a corner cell - the first player can take any cell and ensure they do not lose, but can always force a win if the second player's first move is a "wrong" reply. The optimum strategy for the rest of the play depends on the first reply of the second player.)
*THIS* is what I subscribe for. More videos like this and you'd have reached a million a LONG time ago. Videos that are basically interviews are just, you know, boring. You really should have reached a million a long time ago cause this channel has so much potential. Congrats nonetheless.
Sebastian Elytron This video is kinda just an interview, but whatever.
superscatboy LOL. The bar has been set very low on this channel my friend, here he actually showed us something. Don't get me wrong, it's mostly fascinating content but sometimes I can't help but feel it should just be a podcast lol
Sebastian Elytron Can you give an example of a video that shows us nothing? I must've missed that one.
What's the up with the shirt, is there something interesting about log_2(10)?
so... how did its performance compare to Zuse's machines?
¿Donde puedo descargar el emulador? Descargue uno que parece ser el mismo pero en Windows 10 no se ejecuta por problemas de incompatibilidad.
Apparently the Titan supercomputer only uses 286W at peak. 150kW is a lot more.
Magnetic ring memory, a classic.
The very first computer had a graphical display? Mind blown.
love the shirt!
Came here to ask about that. What's the significance?
Well in the computer world, using logarithms is an easy way to figure out how many bits you need for a an integer. So, for a 10, you need 4 bits. ;-)
Ok. So not a reference to some inside joke, like a Parker square or -1/12.
Maybe there could be, and I just don't know about it. Heh.
what's happened with the rebuild?
How long till the doom port?
I didn't get why they commenting that much on speed of thing, its 70 years old. :) Cycle of operation was 1.5ms (multiplication 6ms) , which means 666 Hz.
Prof: "...roughly a kiloword, 1024 18-bits memory words..."
Me: "Wow, that's a lot!"
Have to admit that I never heard of a DRAM refresh consisting of an oversized thermometer and an intercom system before.
There are 10 kinds of professors on Computerphile:
Those who have Parker cubes in their office
and those who don't
At 12:25, '...in1949, life had not been invented.' threw my literalistic mind into confusion for a moment. ;-D
5:29 What do you mean "Sound travels a lot slower in mercury than in air"? Did you mean in the gaseous form? Because I couldn't even find numbers for that.
A tube with mercury in it. The metal.
Ebumbaya ' sound travels in mercury at rate of 1450m/s, which is 4.22 faster than in air. What professor probably was thinking was that sound travels in mercury slower than in most metals.
No, it's 4.22 times *as fast as* in air, not 4.22 times *faster than* in air. If it was 4.22 times faster than in air, it would be 1790.5 m/s. Because "faster than" implies "on top of what it already is".
You are absolutely correct. Thanks.
Rafal Lasocki maybe he was talking about surface waves? Like on a lake or ocean?
Memory with mercury makes it heavy. What would be the weight of a 64Gb key made with this tech? Several 100d tons?
The mercury Tubes are Separate from the computer not built in. as for size and Weight the computers were already huge and weighed Tons u didnt exactly carry them lol.
So actually the screen shows the content of the memory? Neat, they have invented kind of VRAM. Writing to memory does not interfere with screen sync? Wonder why they did not hook up a reel2reel tape to the audio circuitry running at 38 cm/s, allowing to make a memory dump.
Yes, when I saw that I thought, hey -- they had a memory-mapped bitmap display, quite an advanced feature for the time!
Given the circulating nature of the memory, it probably wasn't very hard to implement, either -- just feed the output from the memory into the CRT along with some timebase signals derived from the clock for scanning.
Writing to the memory wouldn't disturb the display at all, since it has to wait for the right word to come around anyway before overwriting it.
I feel like I want to program my own emulator for this rather than trying to program for it. Does it make sense? I feel like it would be simpler and less painful to write an emulator for a very simple computer using modern languages rather than trying to use obscure machine language to program such computer...
I learned FORTRAN from a graduate student that sarcastically talking down to the the class because we were using “cards” not “paper tape”
I passed the class because “1 +1 = 1.998”
would you like to play a game of thermonuclear war
Hi level is so useful but low level is just so interesting.
Wasn't static ram possible with tubes? Two would make a flip-flop, just like with solid state.
Flip Flops were MUCH Later see computer history.
Mister Hat - It would be possible, but not feasible. The Eccles-Jordan circuit (invented in 1918, incidentally) requires two valves per _bit_ (not per byte or per word), so a 1K memory module would require about 10,000 valves, occupy a large building, and require as much power as a small town.
Why does sound travel slower in mercury than air? I thought that sound was faster in denser mediums.
TheSpacecraftX I think he meant slower in mercury than other metals. Or he got them swapped(air and mercury). Though I could be completely wrong on all counts.
The Mercury enabled the wave to be trapped and Amplified so it could be compared the wave would be slower than in Air Yes.
so I need 3.322 bits for each decimal place in memory?
Yes.
What about magnetic core memory?
Magnetic core memory was not invented until 1955.
I suppose that the Williams tube (used at Machester and disclosed in a 1947 patent) could have been an alternative to the mercury delay line, but had its own set of problems.
Someone made the Tshirt ! :-)
Whats the deal with the log 10 /log 2 = 3.322 Tshirt?
It's how many bits of information there are in a decimal digit.
...or how many octaves there are in an order of magnitude!
can we have the Git Repository so we can build a linux version of it :) ?
Went to the site, and there doesn't seem to be a git repository.
icexiro I think it should run with Wine
Not without FLEX TAPE
Give me the paper tape so I can emulate an edsac on the edsac.
Brant Wedel FLEX TAPE CAN DO ANYTHING
Since temperature has a significant impact on the speed of sound, how did they keep the temperature of those mercury tanks constant? Wouldn't the pretty low specific heat capacity of mercury cause it to heat up very quickly? Sounds like an engineering nightmare to get this right...
Maybe they used some constant 'refresh rate', making the temperature more predictable?
I don't know to be honest.
If it were a problem, I'm sure it wouldn't be hard to arrange a temperature-controlled environment for the mercury.
Alternatively, you could vary the clock rate of the machine according to the temperature of the mercury. :-)
999k subscribers....... Soon, a big day.
bring back rewinding!
But can it run Pong?
Can someone please some that rubicscube?
I wish Prof Brailsford was my grandpa
There is surprisingly little difference between holding words in memory and holding words in mercury.